This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH x86_64][BZ #20033] Use calls to finite scalar versions in vector log, pow, exp.


On Fri, 8 Jul 2016, Andrew Senkevich wrote:

> diff --git a/math/libm-test.inc b/math/libm-test.inc
> index 4ac7a0c..7734333 100644
> --- a/math/libm-test.inc
> +++ b/math/libm-test.inc
> @@ -10631,10 +10631,12 @@ nexttoward_test (void)
> 
>  static const struct test_ff_f_data pow_test_data[] =
>    {
> +#if !(defined __x86_64__ && defined __FAST_MATH__)
>      TEST_ff_f (pow, qnan_value, 0, 1, ERRNO_UNCHANGED),
>      TEST_ff_f (pow, -qnan_value, 0, 1, ERRNO_UNCHANGED),
>      TEST_ff_f (pow, qnan_value, minus_zero, 1, ERRNO_UNCHANGED),
>      TEST_ff_f (pow, -qnan_value, minus_zero, 1, ERRNO_UNCHANGED),
> +#endif

No, such architecture conditionals on individual tests are not 
appropriate.  Maybe you need a NO_TEST_VECTOR handled like NO_TEST_INLINE 
in enable_test.

-- 
Joseph S. Myers
joseph@codesourcery.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]